nano-banana-pro_skill

Overview

This skill generates high-quality images using Google's Nano Banana Pro (Gemini 3 Pro Image) model. It provides ready-to-use patterns and code examples for image creation, iterative editing, and production-grade visuals via the Gemini API. Use it to add professional image generation to apps, marketing workflows, or automated design pipelines.

How this skill works

The skill calls the Gemini 3 Pro Image endpoint to produce images and optional text outputs. It supports direct prompts, inline image references for character consistency, configurable aspect ratios and resolutions, and multi-turn chat sessions for iterative edits. Responses include base64 image data and metadata that you can save, preview, or return from server routes.

When to use it

• Generating production-quality marketing visuals, posters, and infographics.
• Creating consistent character or product shots using reference images.
• Building image generation features in web or mobile apps (APIs or server routes).
• Iteratively editing images via multi-turn conversations (e.g., refine language or composition).
• Embedding text-rendered graphics like event posters or UI mockups.

Best practices

• Store the Gemini API key securely as an environment variable and never hard-code it.
• Request both TEXT and IMAGE modalities to receive captions and metadata alongside images.
• Use image_config (aspect ratio, image_size) to match downstream display or print requirements.
• Provide inline reference images to preserve character consistency across multiple outputs.
• Use chat sessions for iterative edits instead of regenerating from scratch to retain layout context.

Example use cases

• Next.js API route that returns a base64 data URL for client previews.
• Automated generation of product photos with consistent lighting and framing for an ecommerce catalog.
• Creating multilingual infographics by generating a base image and then editing text layers in follow-up messages.
• Design prototyping: produce multiple aspect ratios (1:1, 16:9, 9:16) from one prompt for cross-platform assets.